Artificial IntelligenceITU initiatives
The Summit is THE leading UN platform for global and inclusive dialogue on AI
Hosted by the ITU in partnership with sister UN agencies, XPRIZE Foundation & ACM
THE SUMMIT
Connect AI innovators with problem owners, to identify practical applications of
AI to accelerate progress towards the UN Sustainable Development Goals
Ensure trusted, safe and inclusivedevelopment of AI technologies and
equitable access to their benefits
THE GOAL
2019 SUMMIT IN NUMBERS
SDG4 EducationSDG3 Good
Health and Well
Being SDG10
Human dignity
SDG7
Scaling AI for
Good
AI for Space
2019 BREAKTHROUGH TEAMSThe heart of the Summit…
ITU / WHO Focus Group on
Artificial Intelligence for Health
Topic areas:Cardiovascular disease risk prediction (TG-Cardio)
•Dermatology (TG-Derma)
•Falls among the elderly (TG-Falls)
•Histopathology (TG-Histo)
•Neuro-cognitive diseases (TG-Cogni)
•Outbreak detection (TG-Outbreaks) New
•Ophthalmology (TG-Ophthalmo)
•Psychiatry (TG-Psy)
•Radiotherapy (TG-Radiotherapy)
•Snakebite and snake identification (TG-Snake)
•Symptom assessment (TG-Symptom)
•Tuberculosis (TG-TB)
•Volumetric chest computed tomography (TG-DiagnosticCT)
Key current output documents :
•FG-AI4H Whitepaper
•E-102: Updated call for proposals: use cases, benchmarking,
and data
•D-103: Updated FG-AI4H data acceptance and handling policy
•C-104: Thematic classification scheme
ITU / WHO
Focus
Group on
Artificial
Intelligence
for Health
ITU/WHO Focus Group
AI for HealthFG-AI4H goals:
standardized
framework
for benchmarking and
evaluation
of AI solutions
Artificial Intelligence for
Health (A4IH) offers
substantial
improvements for
public and clinical
health, e.g. early
detection, diagnosis
and risk identification,
treatment decision
support, self-
management,
improved outcomes, …
For world-wide
adoption, need
evaluation standards
on effective AI for
Health
Focus Group AI for
Health (FG-AI4H)
created July 2018;
open platform
AI for Health Use
Case in
Histopathology:
Diagnostic
Support for
Breast Cancer
Treatment
• Tumor infiltrating
lymphocytes (TILs) are
implicated in eliminating
tumor cells
• Quantification of TILs
relevant for patient prognosis
estimation and therapy
selection
• Replace “eye-balling” by
pathologist with Machine
Learning method for TIL
quantification
• Focus Group: specify
process on data generation
and evaluate accuracy of
Machine Learning method
Source: Hendry, S., Salgado, R., Gevaert, T.,
Russell, P. A., John, T., Thapa, B., ... & Sanders, M.
(2017). Assessing Tumor-Infiltrating
Lymphocytes in Solid Tumors: A Practical Review
for Pathologists and Proposal for a
Standardized Method from the International
Immuno-Oncology Biomarkers Working Group
Part 2 (…). Advances in anatomic pathology,
24(6), 311-335.
Copyright 2017 Wolters Kluwer Health, Inc. All
rights reserved.
Unified architecture for machine learning in 5G and future networks processed and approved by SG13 on 1 of July “Architectural framework for machine learning in future networks including IMT-2020"
ITU's ML-Aware Network Architecture: Bringing Intelligence to VerticalsMarch 2019Upcoming: Machine learning in 5G and future networks: use cases and basic requirementsUpcoming: Framework for data handling to enable Machine Learning in future networks including IMT 2020Upcoming: Method for evaluating mobile network intelligence level
ITU-T Focus Group on
Machine Learning for 5G
Source: https://www.itu.int/en/ITU-T/focusgroups/ml5g/Pages/default.aspx
Assessing the economic impact of Artificial Intelligence
Contributed by the McKinsey Global Institute (MGI), the economic and business research arm of McKinsey & Company, this paper offers a framework for thinking about how to model the economic impact of AI
• AI has large potential to contribute to global economic activity.
• The pattern of adoption and full absorption may be relatively rapid—at the high end of what has been observed with other technologies.
• The economic impact may emerge gradually and be visible only over time
• A key challenge is that adoption of AI could widen gaps between countries, companies, and workers.
5AI categories
Computer vision, natural language, virtual assistants, robotic process automation, and advanced machine learning
More countries are taking measures to advance AI (examples - mid 2018)
Economic prosperity
• Lift output and productivity in areas such as agriculture, food production, and other logistics (Goal 2)
• Improve current signaling on tracks through applications of smart sensor technology and advanced analytics (Goal 9)
Social equity
• All students have access to high quality education whatever the student’s inherited circumstances (Goal 4)
• Developing objective and efficient ways to identify and respond to gender bias, discrimination, and violence (Goal 5)
• Monitor users’ financial vitals to improve their financial condition (Goal 10).
Environmental protection
• Improve the efficient management of natural resources and the accountability of harmful activities by using AI (Goal 6).
• Intersection of AI with climate science is assisting researchers to better identify, under-stand, and predict atmospheric processes (Goal 13).
International cooperation
• AI for Good Summit solidifies the UN-wide commitment to partnership and cooperation to scale up AI-enabled innovative solutions to advance the SDGs (Goal 17).
AI and SDG
Augmentation
AI will likely also redefine many existing occupations,
augmenting human capabilities and making
workers more productive
Substitution
The 2017 research conducted by MGI on the impact of automation on
work suggests that roughly half of the time spent on
various tasks could theoretically be automated
by adopting existing technology. The picture could, of course, change
depending on technological progress.
Product and service
innovation and extension
Investment in AI beyond what is needed strictly for
labor substitution can produce additional
economic output by expanding firms’ portfolios,
increasing channels for products and services,
developing new business models, or some
combination of the three
Economic gains from increased
global flows
AI can contribute to digital flows in two ways. The first
is by facilitating more efficient cross-border
commerce.
making improved and expanded use of cross-
border data in flows other than commerce, which can enhance the performance of AI solutions and, in turn,
can improve the productivity of local activities, especially
services
Wealth creation and
reinvestment
higher productivity of economies, the increased
output from efficiency gains and innovations can
be passed to workers in the form of wages and to
entrepreneurs and firms in the form of profits. The
generation of wealth induced by AI could create spillover effects that boost
economic growth.
Transition and implementation
costs
e.g. companies are likely to incur cost restructuring
their organizations, companies might need to pay associated costs such as severance, cover the cost of systems, their
integration, and associated project and consulting fees,
companies also need to build capabilities
Negative externalities
AI could induce major negative distributional externalities affecting workers among others
Production channels External dimensions
7 possible channels of AI impact
AI’s net economic impact -seven dimensions
Automation of labor could add up to about 11 percent or around $9 trillion to global GDP by 2030
Innovation in products and services could deliver up to about 7 percent or around $6 trillion of potential GDP by 2030
Negative externalities and transition costs could reduce the gross GDP impact by about nine percentage points, or around $7 trillion
Micro Drivers: Digitization and Competitive Pressure
Macro Drivers
i) AI investment,ii) AI research activities, iii) Potential productivity boost from AI and automation,iv) Digital absorption,v) Innovation foundation, vi) Human capital, vii) Connectedness, and viii) Labor-market structure and flexibility
Readiness for AI
Gap between country groups is significant and may grow further
Sector analysis
Sector analysis
Sector analysis
Sector analysis
Sector analysis
Thank Youhttps://www.itu.int/en/journal/001/Pages/default.aspx